This tutorial demonstrates some key mapping functionality in R.

1 tmap

1.1 Introduction to tmap

The R package tmap provides an easy way of creating “thematic” maps.

library(tidyverse)
library(tmap)


# A great intro document
# vignette("tmap-getstarted")

#Get tips
tmap_tip()
## An manual legend can be created with tm_add_legend. 
## 
## data(metro)
## tm_shape(metro) + 
##   tm_symbols(col = "pop2020", size = "pop2020", legend.col.show = FALSE, legend.size.show = FALSE) + 
##   tm_add_legend(type = "symbol", border.col = "grey40",
##                 col = tmaptools::get_brewer_pal("YlOrRd", 4, plot = FALSE), 
##                 size = sqrt(seq(1:4) / 4),
##                 labels = c("0 to 10 mln", "10 to 20 mln", "20 to 30 mln", "30 to 40 mln"),
##                 title = "Population in 2020")

There are a number of datasets provided within the tmap package. We will use the “world” data.

data("World")

Inspect the output from the following code.

str(World)
class(World)

# Specifies shape information for each row
World$geometry

The functions tm_shape and tm_polygons are used to create and colour a map.

# Create a blank map
tm_shape(World) + tm_polygons()
# Create a map, colour coded by life expectancy
tm_shape(World) + tm_polygons("life_exp")

Plots can be static or interactive

# Change mode to interactive "view"
tmap_mode("view") 
tm_shape(World) + tm_polygons("life_exp") 
#To toggle between the modes
ttm()

#To switch back to static mode
tmap_mode("plot")

You can merge your own data with World and display it on the graph. Here, we use random numbers to demonstrate the concept.

World$rand <- rnorm(dim(World)[1],0,1)
tm_shape(World) + tm_polygons("rand", midpoint = NA)

1.2 Adding layers to the plot

Multiple layers of data can be added to plots.

data(World, metro, rivers, land)

Explore each part of the code separately to see its function.

tm_shape(land) + 
  tm_raster("elevation", palette = terrain.colors(10))
tm_shape(World) +
  tm_borders("white", lwd = .5) +
  tm_text("iso_a3", size = "AREA") 
tm_shape(metro) +
    tm_symbols(col = "red", size = "pop2020", scale = .5)

These commands can be combined as follows. Notice how the layers get overlaid.

tm_shape(land) + 
  tm_raster("elevation", palette = terrain.colors(10)) +
tm_shape(World) +
  tm_borders("white", lwd = .5) +
  tm_text("iso_a3", size = "AREA") +
tm_shape(metro) +
    tm_symbols(col = "red", size = "pop2020", scale = .5) +
tm_legend(show = FALSE)

1.3 Customising the map

Colours, titles and margins etc can be customised.

Options can also be set locally using tm_layout.

tm_shape(World) + tm_polygons("life_exp") +
tm_layout(bg.color = "skyblue", inner.margins = c(0, .02, .02, .02))

Options can also be set globally using tmap_options and tmap_style.

tmap_options(bg.color = "black", legend.text.color = "white")
tm_shape(World) + tm_polygons("life_exp")


tmap_style("classic")
tm_shape(World) + tm_polygons("life_exp")  
tmap_style("watercolor")

1.4 Exporting maps

tm <- tm_shape(World) + tm_polygons("life_exp")  

## save an image ("plot" mode)
tmap_save(tm, filename = "world_map.png")

## save as stand-alone HTML file ("view" mode)
tmap_save(tm, filename = "world_map.html")

1.5 Exercises

Start from this plot:

tmap_mode("plot")
tm_shape(World) + tm_polygons("life_exp")

  • Use tm_layout to:
  • Add a title in the top right corner of the map
  • Add a title above the map
  • Change the background colour to a colour of your choice
  • Add an option to tm_polygons to change the colour palette
  • Change the style of the plot to “watercolor”
rm(list=ls())

2 ggmap

2.1 Introduction to ggmap

The package ggmap uses the conventions of ggplot2 and applies them to mapping.

Load the package.

library(ggmap)

In ggmap (and many other mapping packages), there are two step: 1) get the map data and 2) create the plot. The function qmap perfoms both steps at once.

2.2 qmap

The function qmap (short for quick map) allows you to create a map very quickly.

qmap(location="auckland")

We can be even more specific with our location. The location string can be anything that you would type into google.

qmap(location="AUT university", zoom=17, maptype = "satellite")

We will now add a marker our current location. First we will get the coordinates of our current location with the geocode function.

wz <- "WZ building, AUT university"
wz_gc <- geocode(wz)
wz_gc
##        lon       lat
## 1 174.7668 -36.85404

Now we will create a map.

g <- qmap(location="AUT, NZ", maprange = TRUE, zoom=18,
          maptype = "satellite", 
          base_layer = ggplot(aes(x=lon, y=lat), data = wz_gc )) 
## Warning: `panel.margin` is deprecated. Please use `panel.spacing` property
## instead
g + geom_point(colour="purple", size=5)

Note: since these functions are querying google, there are some limitations in terms of maximum queries per day (2500) and per second. You may get an “OVER_QUERY_LIMIT” message. Waiting a few seconds should resolve the issue. You can check your daily limit as follows.

geocodeQueryCheck()

2.3 get_map and ggmap

The get_map function is a wrapper for assorted get_* functions. Once the map has been obtained using get_map, it can be plotted using ggmap.

m <- get_map(wz_gc, zoom=14)
g <- ggmap(m)
g

The get_map function allows different map sources.

m <- get_map(wz_gc, zoom=14, source = "stamen", maptype = "watercolor")
g <- ggmap(m)
g

m <- get_map(wz_gc, zoom=14, source="google", maptype = "roadmap")
g <- ggmap(m)
g

Run the following code and notice how different “extent” options change the look of the map.

ggmap(m, extent="device")
ggmap(m, extent="normal")
ggmap(m, extent="panel")

2.4 Adding paths to a map

Paths (pairs of longitudes and latitudes) can be added to plots.

# Create a random path
set.seed(500)
df <- round(data.frame(
x = jitter(rep(wz_gc$lon, 10), amount = .01),
y = jitter(rep(wz_gc$lat, 10), amount = .01)
), digits = 3)

# Create the map
map <- get_googlemap("St Pauls street, Auckland", markers = df, path = df, scale = 2, zoom=14)

# Plot the map
f <- ggmap(map, extent = 'device')
## Warning: `panel.margin` is deprecated. Please use `panel.spacing` property
## instead
f

Let us inspect the map object.

str(map)
##  'ggmap' chr [1:1280, 1:1280] "#AADAFF" "#AADAFF" "#AADAFF" "#AADAFF" ...
##  - attr(*, "bb")='data.frame':   1 obs. of  4 variables:
##   ..$ ll.lat: num -36.9
##   ..$ ll.lon: num 175
##   ..$ ur.lat: num -36.8
##   ..$ ur.lon: num 175
##  - attr(*, "source")= chr "google"
##  - attr(*, "maptype")= chr "terrain"
##  - attr(*, "zoom")= num 14
attr(map, "bb")
##      ll.lat  ll.lon    ur.lat   ur.lon
## 1 -36.87587 174.739 -36.83192 174.7939

We can colour code by a third variable.

First we load the data.

# Dummy data
citywalk <- data.frame(
  lon = c(174.766, 174.767, 174.768, 174.769),
  lat = c(-36.853, -36.854, -36.855, -36.856),
  elevation = c(1,2,3,4)
)

Then we create the map.

citymap <- get_map(location = c(mean(citywalk$lon), mean(citywalk$lat)), 
                   maptype = "roadmap", 
                   source = "google", 
                   zoom = 14)
ggmap(citymap) + 
  geom_path(data = citywalk, aes(color = elevation), size = 3, lineend = "round") + 
  scale_color_gradientn(colours = rainbow(7), breaks = seq(0, 200, by = 25))

2.5 Exercises

  • The file “route_210183.csv” contains a path around Auckland. Plot the path over an Auckland map and determine the highest point on the map?
  • Create a map of the North Island. Add a path between Auckland and Wellington, going via Taupo.
rm(list=ls())

3 Leaflet

The final R package we will explore is leaflet.

library(leaflet)
library(rgdal)
library(rgeos)

We can create a map of our current location.

aut <- data.frame(ID = c("AUT"),
                     x = c(174.7664924),
                     y = c(-36.8539873), 
                     group=NA)

m <- leaflet() %>%
     addTiles() %>% # Add default OpenStreetMap map tiles  
     setView(lng = aut$x, lat = aut$y, zoom = 15) 

m  # Print the map

Try these functions

m %>% addMarkers(lng= aut$x, lat = aut$y, popup="WZ building") 

m  %>% addPopups(lng= aut$x, lat = aut$y, "Here is the <b>WZ building</b>, AUT")


rand_lng <- function(n = 10) rnorm(n, aut$x, .01)
rand_lat <- function(n = 10) rnorm(n, aut$y, .01)

# circle (units in metres)
m %>% addCircles(rand_lng(10), rand_lat(10), radius = runif(10, 50, 150))

# polyline
m %>% addPolylines(rand_lng(10), rand_lat(10))
rm(list=ls())

4 Solutions

4.1 tmap

library(tmap)
data("World")
tmap_mode("plot")
tmap_style("watercolor")
tm_shape(World) + tm_polygons("life_exp",  palette = "RdBu")   + 
  tm_layout(bg.color = "purple", title="A title", main.title="Another title")

4.2 ggmap

library(ggmap)
library(readr)
library(tidyverse)
citywalk <- read_csv("data/route_210183.csv", skip=1) # data from MapMyRun.com
citywalk <- dplyr::rename(citywalk, elevation=`elevation(meters)`)
citywalk$lat  <-as.numeric(citywalk$lat)
citymap <- get_map(location = c(mean(citywalk$lon), mean(citywalk$lat)), 
                   maptype = "roadmap", 
                   source = "google", 
                   zoom = 14)
ggmap(citymap) + 
  geom_path(data = citywalk, aes(color = elevation), size = 3, lineend = "round") + 
  scale_color_gradientn(colours = rainbow(7), breaks = seq(0, 200, by = 25))

locations <- data.frame(city = c("Auckland", "Taupo", "Wellington"), stringsAsFactors =FALSE)

# Get long & lat for locations
# Option 1: using geocode
# locations <- cbind(geocode(locations$city), locations)
# Option 2: using data from elsewhere
locations$lon <- c(174.7633, 176.0702, 174.7762)
locations$lat <- c(-36.84846, -38.68569, -41.28646)


summary(locations)
##      city                lon             lat        
##  Length:3           Min.   :174.8   Min.   :-41.29  
##  Class :character   1st Qu.:174.8   1st Qu.:-39.99  
##  Mode  :character   Median :174.8   Median :-38.69  
##                     Mean   :175.2   Mean   :-38.94  
##                     3rd Qu.:175.4   3rd Qu.:-37.77  
##                     Max.   :176.1   Max.   :-36.85
# Create the map
map <- get_googlemap("North Island, New Zealand", 
               markers = locations[,c("lon", "lat")], 
               path = locations[,c("lon", "lat")], 
               scale = 2, zoom=6)

# Plot the map
f <- ggmap(map)
f

5 Resources

Good overview of mapping tools https://cran.r-project.org/doc/contrib/intro-spatial-rl.pdf

https://rpubs.com/nickbearman/r-google-map-making

Good introduction to the ggmap package http://stat405.had.co.nz/ggmap.pdf

More advanced information about geo-data https://geocompr.robinlovelace.net/adv-map.html

Useful overview of key R packages https://www.computerworld.com/article/2921176/business-intelligence/great-r-packages-for-data-import-wrangling-visualization.html

6 Acknowledgements & References